[Mpi-forum] Discussion points from the MPI-<next> discussion today

[Underwood, Keith D]

> > Passive one-sided communication, for a start.  I no longer have access
> > to the sort of systems on which that would fail, and I have never seen
> > it used.  When I did have such access, I did write code that caused
> > its underlying assumptions to break, while investigating problems on
> > the systems I managed.
> 
> You still haven't said what the actual problem is.  I suspect that you just
> weren't using it properly if you broke the assumptions of MPI-2.2 RMA.  It
> was extremely conservative.

I have to agree.  It seems much, much more likely that you found a bug in the way a given implementation interacts with whatever you were doing with C or that your usage of RMA was not technically conformant.  That said, if you have (or could reconstruct) an example that breaks the assumptions of MPI-2 RMA, it would be very beneficial to the standard if you would post that to the list.  If the model is broken, we need to fix it!

Keith